AI 捏捏特效 / 挤压效果万相视频 LoRA，是由专业团队精心研发，基于阿里云开源的万相 2.1（Wan）14B 参数规模的 I2V（图像到视频）480p 模型开展训练工作。该模型在权威评测集 VBench 中成绩优异，以总分 86.22% 的成绩登顶全球榜首。在此基础上训练出的 LoRA，能为视频增添独特的挤压特效，让画面呈现出别样的动态效果。
14B I2V 480p v1.0：
使用 1.5 分钟的视频进行训练，该视频由 20 个物体被挤压的短片组成。这是在 Wan.21 14B I2V 480p 模型上进行训练的。
触发词是：’sq41sh squish effect‘
请参阅下面的一些示例提示，它们效果很好。我相信提示结构与示例非常相似；您只需要指定被挤压的对象，这就是唯一的区别。
推荐设置：
LoRA强度=1.0
Embedded guidance scale = 6.0
Flow shift = 5.0
这是我用来生成这些视频的 Wan2.1 I2V 推理工作流程的链接：https://github.com/kijai/ComfyUI-WanVideoWrapper/blob/main/example_workflows/wanvideo_480p_I2V_example_02.json
提示示例：
示例 1： In the video, a miniature rodent is presented. The rodent is held in a person’s hands. The person then presses on the rodent, causing a sq41sh squish effect. The person keeps pressing down on the rodent, further showing the sq41sh squish effect.
示例 2：In the video, a miniature tank is presented. The tank is held in a person’s hands. The person then presses on the tank, causing a sq41sh squish effect. The person keeps pressing down on the tank, further showing the sq41sh squish effect.
如您所见，提示结构是相同的，只有被挤压的物体发生了变化。
| 触发词： | sq41sh squish effect |
| 使用提示： | Strength: 1 |
作品参照
In the video, a miniature rodent is presented. The rodent is held in a person’s hands. The person then presses on the rodent, causing a sq41sh squish effect. The person keeps pressing down on the rodent, further showing the sq41sh squish effect.
Steps: 20, CFG scale: 6, Sampler: DPM++ SDEIn the video, LeBron James is presented. LeBron James is held in a person’s hands. The person then presses on LeBron James, causing a sq41sh squish effect. The person keeps pressing down on LeBron James, further showing the sq41sh squish effect.
Negative prompt: Vibrant colors, overexposed, static, blurry details, subtitles, style, artwork, painting, image, still, overall grayish, worst quality, low quality, JPEG compression artifacts, ugly, incomplete, extra fingers, poorly drawn hands, poorly drawn face, deformed, disfigured, deformed limbs, merged fingers, static image, messy background, three legs, many people in background, walking backwards, bad quality video
Steps: 20, CFG scale: 6
